Structural FOUNDATIONS Pile Load Testing for Bored Piles in Soil a Brief Summary by Hee Yang Ng, Mistructe, C.Eng, P.E

structural FOUNDATIONS Pile Load Testing for Bored Piles in Soil A Brief Summary By Hee Yang Ng, MIStructE, C.Eng, P.E.

ored piles, also referred to as auger-cast piles, are large-diameter at intervals throughout the length of the pile so that the force for cast-in-place concrete piles used commonly to support buildings each segment can be calculated and the pile’s load transfer can be Bwhen column loads are high. The design of bored piles requires the obtained. The same pile installation method should be used to install designer or the project’s geotechnical engineer to the working piles for the project. To adequately estimate the load-carrying capacity of the pile based verify the shaft friction and end-bearing capacity, on the ground conditions at the project site. One preliminary piles are typically loaded to 3 times way to predict the pile capacity is by carrying out the working load. calculations using soil parameters obtained from Working pile tests are conducted during piling site investigation data. While many pile design work to verify the quality and workmanship of methods are highly empirical, pile capacity can also the installed piles. These tests verify that installed be affected by other factors such as the method of piles perform as-designed to meet strength (abil- installation, quality of workmanship, and construc- ity to carry load) and serviceability criteria (with tion materials used. Therefore, pile load testing is acceptable settlement). Working piles are loaded an essential aspect of pile design that should not progressively, with the load and pile head settle- be overlooked. ment monitored and recorded. The acceptance This article summarizes pile load testing (static criteria for working piles are settlement limits at type using a kentledge) for bored piles in soil and Figure 1. Generalized load-settlement certain loads. For example, in CP4 (SS CP4 Code highlights some key aspects designers should look curve of a pile. of practice for foundations, Singapore Standard), for when reviewing and interpreting pile load test the settlement cannot exceed 0.59 inches (15mm) results. For more information on pile design, the at 1.5 times the working load or 0.98 inches reader is directed to Pile Structural Capacity – A (25mm) at 2 times the working load. The higher Comparison of Three Design Codes published in the allowable value of settlement at 2 times of work- March 2020 issue of STRUCTURE. ing load is to recognize that, at higher loads, the rate of increase in settlement is likely to be greater. Recommendations for the number of tests are Types of Load Tests shown in Table 1. There are essentially two main types of pile load tests: the preliminary pile test and the working pile test. A preliminary pile test is typically carried Load-Settlement Curve out before the piling works commence while the A simplified load-settlement curve is shown in Figure 1. working pile test is carried out during piling or Generally, the curve is non-linear, and the stiff- when substantial piling works are completed. When ness of the response tends to degrade or soften as using limit state design (e.g., Eurocode 7-EC7), the the load gets higher. This is expected because, as preliminary pile test and working pile test allow the load increases, the pile approaches “failure” in designers to adopt less conservative partial factors Figure 2. The difference in mobilization geotechnical terms. Some codes define the “failure” of shaft friction and end-bearing. when reducing unfactored shaft friction and end- of a pile as reaching a settlement of 10% of pile bearing, resulting in a more economical design. Similarly, when using diameter. For large diameter piles, this criterion becomes somewhat LRFD design or ASSHTO specifications, more load testing may allow lenient if the serviceability limit state is considered. For example, a a higher resistance factor (φ) to be used due to the increased reliability. 39.37-inch-diameter (1000mm) bored pile has an allowable settle- The preliminary pile test verifies the design parameters prior to the ment of 3.94 inches (100mm). The settlement acceptance criterion start of piling work. Therefore, the test pile is instrumented so that has to be stringent because pile settlement can result in a differential assumed values of shaft friction and end-bearing could be verified. A settlement between supports, which is a grave concern due to the borehole close to the pile location is necessary to ascertain the type amplification of tilt for a tall superstructure. In addition, build- and layering of the soil for verification. Strain gauges are installed ing settlements may disrupt utilities and services (e.g., pipes and cables). Therefore, a much smaller value of the allowable settlement Table 1. Number of pile load tests. is usually required. Piles rarely “fail” with regard to vertical structural capacity. In a case Preliminary pile test 1 pile or 0.5% of the total number of working piles, whichever is greater. study presented later, a quick calculation shows that the axial stress in the pile at its rated working load is on the order of about 8 to 10 Working pile test 2 piles or 1% of the total number of work- MPa (or 1.2 to 1.5 ksi). ing piles or 1 for every 50 meters length of A critical aspect of pile design is shown in Figure 2, where the proposed building, whichever is greater. load on a pile installed in soil is resisted by a combination of shaft

52 STRUCTURE magazine friction and end-bearing. As the pile is loaded, shaft friction increases pile with its load-settlement path running along the linear-elastic gradually and can become fully developed by a small movement portion of the curve will likely return to zero load with minimal of approximately 0.5% of pile diameter, say residual settlement. 0.2 to 0.4 inches (5mm to 10mm). However, Sometimes it can be helpful to compare the gradient a large movement is required for end-bearing of the load-settlement curve against the elastic short- to be fully developed, typically in the range of ening gradient (PL/AE line). For example, the elastic 5% of pile diameter. This means that to achieve shortening of a bored pile without any confinement its full end-bearing value, the pile must settle is given by PL/AE, where P is the applied load, L is more, which may not be desirable. For this the length of pile, A is the cross-sectional area, and E reason, designers tend to select more conserva- is Young’s Modulus. tive values for end-bearing, and codes impose The load-settlement gradient of a friction pile a higher factor of safety. should generally be similar or even very slightly Figure 3 shows some examples of load-settle- steeper compared to the elastic shortening gradient. ment curves of piles with anomalies compared This is because of the presence of soil confinement to a pile expected to perform normally. The and shaft resistance. However, suppose the gradient red curve is abnormal because of kinks in the of the load-settlement of a pile is gentler than the curve, possibly resulting from pile defects or Figure 3. Examples of “abnormal” elastic shortening gradient. In that case, it may be load-settlement curve of a pile. unexpected ground conditions that cause the a case of less friction being developed (e.g., softer pile to settle abruptly when loaded. This inability to sustain load at soils) along the shaft or a case of stiffer soils located only at the localized portions of the curve is not acceptable. The purple curve lower end of the pile. shows normalcy up to a point and degrades abruptly, finding a plateau with a lower load and increasing settlement. This might occur if the pile base is poor or suddenly loses support coupled Calculation of Pile Settlement with a loss of a portion of shaft friction or deterioration of shaft There are various methods to estimate anticipated pile settlement in friction. One example is a case of cavities, archetypal in limestone design. One of the easiest ways to calculate pile settlement by hand is areas. Lastly, the blue curve is abnormal due to its inability to reach by Vesic (1977), as shown in Figure 5 (page 54). In this method, the the expected design load and experiences excessive settlement too pile head settlement is the sum of three components, namely axial early. This might occur for a pile that has fully maximized the shaft compression (ws), the settlement at pile toe due to shaft load (wps), friction and cannot develop the end-bearing, resulting in a “plung- and settlement at pile toe due to end-bearing load (wpp). ing” type of failure. Axial compression in soil can be simplified by assuming it as 75% of unconfined elastic shortening (PL/AE). The formula given by Vesic uses the sum of end-bearing load and a fraction (0.5 to 0.67) Case Study of shaft friction load in place of axial load in the unconfined elastic A 31.5-inch-diameter (800mm) bored pile was installed to a depth shortening formula to calculate the axial shortening of the pile in of 85.3 feet (26m) below ground in alluvial soil. These soils were soil. Settlements due to shaft load and end-bearing load are given by formed by transported material deposited and cemented over time. the ratio of shaft design working load (Qs) and end-bearing design The working load for such a pile is 450 tons (4000 kN). Please note working load (Qb) against the ultimate bearing capacity at the pile that pile capacities provided by geotechnical engineers are always toe (Qo), normalized by the relevant dimension (pile length L and stated as allowable loads (ASD) or working loads, typically with a pile diameter d, respectively) and multiplied by Cp and Cs, empirical safety factor of two against a geotechnical failure. coefficients subject to variability. Because of this, it is appropriate to Figure 4 shows the load-settlement curve for an instrumented pre- study the sensitivity to this variability during the design process and liminary pile test carried out for the 31.5-inch-diameter (800mm) benchmark against past experience in similar ground for similar piles bored pile. The pile was loaded to 3 times the working load so the before relying too much on the result. full value for shaft friction could be developed. During the applica- What is important to note in using this method is that the value tion of the lower range loads (red line), the curve is approximately of settlement calculated is highly dependent on what value of Cp is linear and somewhat elastic. This means that the pile is behaving selected and the value of ultimate base capacity qo. The method recog- well, and the response has not reached a state of yielding. When nizes that sand normally has a higher ultimate bearing capacity than loads are higher (green line), it can be seen that the gradient of the curve has decreased, meaning stiffness has degraded. This behavior is analogous to materials approaching yield when a higher load range increases settlement. After reaching 3 times the design working load, the pile was unloaded (purple line). The unloading line usually has a steeper gradient because the pile loading process stiffens the soil around the pile. Lastly, the final portion of unloading to zero load (blue line) shows a very slight stiffness reduction similar to reaching yield during the loading process. However, this may not be obvious or observed in all tests. After unloading, it is helpful to examine the residual settlement (0.43 inches [11mm] in Figure 4). This is the settlement that is per- manent and cannot be recovered. A high value of residual settlement usually means that the pile has been loaded beyond yield, result- ing in large permanent deformation. Conversely, a lightly loaded Figure 4. Preliminary pile test result.

JUNE 2021 53 clay. For example, the limit on end-bearing shaft friction is reached, the rest of the load can be 40,000 kPa compared to 15,000 kPa would be taken by increasing end-bearing, for sand and clay, respectively. Therefore, the which comes at the expense of a much higher Cp value for clay is correspondingly lower com- settlement. For loads up to the working load, pared to sand. as the design of the pile has already taken into It can be noted that most pile settlement consideration developing very little end-bear- comes from axial compression and base set- ing and mainly relying on shaft friction, Vesic’s tlement. Because of this, designers who need equation predicts settlement quite accurately, as to control settlement might want to consider seen in this example, when using an artificially increasing pile stiffness (e.g., higher grade con- high qo to minimize the contribution from crete and increased steel reinforcement area) component wpp. and improving the pile toe condition (e.g., base However, when using a value of qo = 3,000 grouting). Another strategy is to design the pile kN/m2, which is the actual value used in the as a pure friction pile (i.e., there is no reliance pile design, Vesic’s equation over-predicts settle- on end-bearing to achieve the pile capacity). ment significantly. Therefore, designers need In such an instance, it is possible to reduce the to be cautioned on the sensitivity of qo and overall safety factor on shaft friction to 1.3. notice that qo may also vary due to size effects. From Figure 6, it can be observed that the pile This illustration is not meant to be an attempt settlement at working load is approximately to validate or invalidate Vesic’s equation for equal to the unconfined elastic compression any case. Designers still need to consider the (i.e., PL/AE). This can be a very crude approxi- Figure 5. Pile settlement calculation by Vesic (1977). design assumptions, ground conditions, past mation to estimate pile settlement without any experience, sensitivity, etc., when calibrating further computation. However, for a friction-only pile, the settlement this equation for use in a new situation. should be halved. This assumes the settlement due to shaft load is small and there is zero settlement due to end-bearing load. Thus, the only component left is the axial pile compression which reduces to Other Points to Note about 0.5PL/AE, according to Vesic’s equation. In loading a test pile, a kentledge reaction system (stacking up concrete Figure 6 shows a comparison of the load settlement curve predicted blocks or steel plates as a counterweight) is customarily used. However, using Vesic’s equation against data from an ultimate load test for the the process is slow, requires significant space, and is not practical for same 31.5-inch-diameter (800mm) bored pile with a working load very high loadings. Nonetheless, many designers prefer this method due of 4,000 kN. It can be seen that the Vesic equation predicted the pile to the easy interpretation of test results. The stacking of high loads on 2 settlement quite well (by assuming qo = 40,000 kN/m ), but only up to a small area requires the ground bearing capacity to be checked. There the working load of 4,000 kN (at about 0.2-inch [5mm] settlement). have been instances of kentledge collapses due to inadequate bearing Beyond that, the predicted settlement is much higher than the actual capacity. Sometimes, it is not practical to test a large diameter pile due settlement. The reason for this is the conservative assumption of to the high loads. Some codes (e.g., Eurocode 7-EC7) allow results of a maximum shaft friction. Using Vesic’s equation, once the maximum pile load test to be extrapolated for a pile diameter not exceeding 2 times.

Conclusion For buildings with high column loads, bored piles are often adopted. However, due to the customary practice and compliance with regulatory requirements, pile load tests are almost always required. The provision of pile load tests allows the designer to verify design parameters, design more economical piles, check the quality of materials and workmanship, and ultimately show that the performance meets load-settlement requirements. It is important to understand the difference in development rate for shaft friction and end-bearing. The load-settlement curve is an important plot to assist designers in understanding the pile behavior. Pile settlement can be easily calculated using Vesic’s equations but, just like any settlement calculations, the value cannot be taken to be precise. It is only a means to offer insights to the designer as a basis for comparison when carrying out a pile load test. It is essential to scrutinize the load-displacement curve of a pile-load test to be satisfied that the behavior is within expectation.■

References are included in the PDF version of the article at STRUCTUREmag.org.

Hee Yang Ng is a Principal Engineer with a building control agency in the Asia-Pacific region. Figure 6. Comparison of load settlement curve.

54 STRUCTURE magazine References ACI-318-14, Building Code Requirements for Structural Concrete and Commentary ACI-336.3R-94, Design and Construction of Drilled Piers Singapore Standard SS CP4:2003, “Code of Practice for Foundation,” SPRING Singapore, ISBN 9971-67-914-0 Eurocode EN 1992-1-1:2004, Design of Concrete Structures. General rules and rules for buildings (commonly known as Eurocode 2 or EC2) Eurocode EN 1997-1, Geotechnical Design – General Rules (commonly known as Eurocode 7 or EC7) Eurocode EN 1536:2010, Execution of Special Geotechnical Works – Bored Piles Vesic, A. S. (1977) Design of Pile Foundations, National Cooperative Highway Research Program, Synthesis of Practice No. 42, Transportation Research Board, Washington D.C.

JUNE 2021 55